最小工作时长约束下的传感器调度方法  

Sensor Scheduling Method under Minimum Working Timestep Constrained

作　　者：安雷 李召瑞 吉兵 陈志刚[2] AN Lei;LI Zhaorui;JI Bing;CHEN Zhigang(Shijiazhuang Campus,Army Engineering University,Shijiazhuang 050003,China;Unit 32272 of PLA,Lanzhou 733000,China)

机构地区：[1]陆军工程大学石家庄校区,河北石家庄050003 [2]中国人民解放军32272部队,甘肃兰州733000

出　　处：《探测与控制学报》2022年第4期56-65,共10页Journal of Detection & Control

基　　金：国防预研项目(41404030101,41404030102)。

摘　　要：针对多传感器目标跟踪中系统频繁切换的问题,基于最小工作时长思想,提出一种切换代价控制的传感器调度方法。首先依据部分可观马尔可夫决策过程确立调度模型,创建跟踪精度约束下的目标优化函数;随后将交互式多模型算法与模糊积分粒子滤波算法相结合,进行目标状态估计,并以后验克拉美罗下界作为衡量指标,实现长时目标跟踪精度的预测;最后以跟踪精度预测结果为依据,结合最小工作时长约束策略,提出调度方法及实现流程。仿真结果表明,该方法在满足目标跟踪精度要求的前提下,进一步降低了传感器切换次数,有效提高了多传感器系统目标跟踪的稳定性和有效性。For theproblem of frequentsystem switching in multi-sensor target tracking,a sensor scheduling method under the switching cost control was proposed based on the idea of minimum working timestep.Firstly,the scheduling model was established according to the partially observable Markov decision process,and the objective optimization function under the constraint of tracking accuracy was created;Then,the interacting multiple model algorithm was combined with the fuzzy quadrature particle filter algorithm to estimate the target state,and the posterior Cramér-Rao lower bound was used as the measurement index to realize the prediction of non-myopic target tracking accuracy;Finally,based on the prediction results of tracking accuracy and combined with the minimum working timestep constraint strategy,the scheduling method and implementation process were proposed.Simulation results showed that the proposed scheduling method further reduced the sensor switching times and effectively improved the stability and effectiveness of target tracking in multi-sensor system on the premise of target tracking accuracy.

关 键 词：传感器调度 最小时长约束 切换代价 目标跟踪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TN95[自动化与计算机技术—计算机科学与技术]